PAIRS AutoGeo: an Automated Machine Learning Framework for Massive Geospatial Data
This framework addresses the challenge of leveraging machine learning for geospatial data for users without extensive geospatial expertise, simplifying the development of industrial solutions.
This paper introduces PAIRS AutoGeo, an automated machine learning framework for geospatial data on the IBM PAIRS Geoscope platform. It simplifies the development of ML solutions by minimizing user input to labeled GPS coordinates, automatically gathering and assembling data, performing quality checks, and training models. For a 10-way tree species classification task, it achieved 59.8% accuracy with a random forest classifier and 81.4% with a modified ResNet model.
An automated machine learning framework for geospatial data named PAIRS AutoGeo is introduced on IBM PAIRS Geoscope big data and analytics platform. The framework simplifies the development of industrial machine learning solutions leveraging geospatial data to the extent that the user inputs are minimized to merely a text file containing labeled GPS coordinates. PAIRS AutoGeo automatically gathers required data at the location coordinates, assembles the training data, performs quality check, and trains multiple machine learning models for subsequent deployment. The framework is validated using a realistic industrial use case of tree species classification. Open-source tree species data are used as the input to train a random forest classifier and a modified ResNet model for 10-way tree species classification based on aerial imagery, which leads to an accuracy of $59.8\%$ and $81.4\%$, respectively. This use case exemplifies how PAIRS AutoGeo enables users to leverage machine learning without extensive geospatial expertise.